AI Glossary: 30+ Voice AI & Business Automation Terms Explained

AI Agent

Software that can perceive its environment, make decisions, and take actions autonomously to accomplish a goal. Unlike scripted chatbots, an AI agent reasons about context, uses tools (calendar lookups, database queries), and adapts in real time.

AI Voice Agent

An AI agent that communicates through spoken language over phone calls or voice interfaces. Combines speech recognition, natural language understanding, and text-to-speech to hold real-time voice conversations.

ASR

Technology that converts spoken words into text. The first step in any voice AI pipeline: the caller speaks, ASR transcribes those words, and the AI processes the resulting text. Source

Chatbot

Software that conducts text-based conversations, typically on websites, messaging apps, or social media. Ranges from simple rule-based systems to AI-powered assistants that generate dynamic responses.

CRM

A system that stores and organizes customer data: contact information, interaction history, purchase records, and notes. When integrated with AI, the CRM becomes a live context source for personalized conversations.

Customer Memory

An AI system's ability to recall information from previous interactions with the same customer, allowing the AI to greet returning callers by name and reference past conversations.

Digital Administrator

An AI system that handles routine administrative tasks typically performed by a front-desk receptionist: answering phone calls, scheduling appointments, responding to common questions, and routing complex requests.

E-E-A-T

A framework Google uses to evaluate content quality for search rankings. Signals whether a website's content reflects genuine real-world experience and credible expertise. Source

GDPR

The European Union regulation governing how organizations collect, store, and process personal data. Requires transparent data usage policies, deletion rights, secure storage, and clear consent mechanisms. Source

Hallucination

When an AI system generates confident-sounding information that is factually incorrect or fabricated. A known limitation of large language models, mitigated by grounding responses in a verified knowledge base.

Intent Recognition

The AI's ability to determine what a caller wants to accomplish from their utterance, identifying both the intent (e.g. reschedule) and relevant entities (e.g. next Tuesday).

IVR

The traditional phone menu system: 'Press 1 for sales, press 2 for support.' Routes callers through fixed decision trees using touch-tone or simple voice commands. Source

Knowledge Base

The structured collection of information an AI system draws from when answering questions: services, pricing, hours, location, policies, and FAQs. Quality of the knowledge base directly determines quality of responses.

LLM

The AI engine behind modern conversational systems. Models like GPT-4, Claude, and Gemini are trained on vast text data to understand and generate human-like language. Source

Latency

The delay between a caller finishing their sentence and the AI beginning its response. For natural voice conversation, sub-500ms end-to-end latency is the target.

Multi-turn Conversation

A dialogue that spans multiple exchanges where the AI maintains context across the entire conversation, remembering earlier statements and building on established information.

NLP

The field of AI focused on enabling computers to understand, interpret, and generate human language. Encompasses intent recognition, sentiment analysis, entity extraction, and language generation. Source

Omnichannel

A unified customer service approach across all communication channels (phone, email, chat, SMS, social, in-person) where context follows the customer from one channel to another.

PMS

Core software hotels use to manage reservations, room inventory, guest profiles, billing, and housekeeping. Examples: Opera, Mews, Cloudbeds, Little Hotelier.

Prompt

The instruction or context given to an AI model to guide its behavior. Defines the agent's personality, knowledge boundaries, rules, and workflows.

RAG

A technique that improves AI accuracy by combining a language model with a search system that retrieves relevant documents from a knowledge base in real time before generating a response. Source

Real-time API

A programming interface that processes data with minimal latency to enable instant interaction. In voice AI, allows speech to be streamed directly to and from the AI model, eliminating batch-processing delays.

SIP

The standard protocol used to initiate, maintain, and terminate voice calls over the internet (VoIP). Defined in IETF RFC 3261. Source

STT

The process of converting spoken audio into written text. The input layer of any voice AI system. Also called ASR. Source

TTS

Technology that converts written text into spoken audio. The output layer of a voice AI system: the AI generates a text response and TTS converts it into natural-sounding speech. Source

Token

The basic unit of text that language models process. A token can be a word, part of a word, or punctuation. Determines context-window size and how AI usage is billed.

Transfer (Call)

When the AI hands an active phone call to a human agent. A warm transfer also passes along a conversation summary so the human does not need to start from scratch.

Voice AI

Umbrella term for AI systems that interact through spoken language. Encompasses the full stack: speech recognition (ASR/STT), natural language understanding (NLP), response generation (LLM), and speech synthesis (TTS).

Voice Widget

An embeddable interface that lets website visitors interact with an AI voice agent directly from a web page using their microphone, with spoken responses.

Webhook

An automated HTTP message sent from one system to another when a specific event occurs. Used to trigger post-call actions like emails, CRM record creation, or calendar updates. Source

Frequently Asked Questions

Related Articles